The bulk of evolutionary ecology implicitly assumes that ecology shapes evolution, rather than vice versa, but there is increasing interest in the possibility of a two-way interaction. Dynamic feedbacks between ecological and evolutionary processes (eco-evo feedbacks) have long been recognized in the theoretical literature, and the observation of rapid evolution has since inspired empiricists to explore the consequences of these feedbacks. Laboratory studies prove that short-term evolutionary change can significantly alter ecological dynamics, particularly in pair-wise interactions. We know far less about whether these reciprocal dynamics are important in more complex natural systems. Here, we outline our approach to that question, focusing on the Trinidadian guppy and the stream ecosystems it inhabits. We summarize results from several types of studies: comparative demography in two types of communities, experiments in mesocosms, common garden laboratory experiments and replicated introduction experiments. The latter were designed as perturbations to the natural steady state that allow us to follow the joint ecological and evolutionary dynamics of guppies and their ecosystem. In each approach, we replicated experiments across multiple independent origins of guppy population types and found that eco-evo feedbacks play major roles in guppy evolution. There are three possible sources for these feedbacks, all of which have some support in our data, which will form the focus of future research efforts.